Modeling Turbulent Flows Introductory FLUENT Training © 2006 ANSYS, Inc. All rights reserved. ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com What is Turbulence? Unsteady, irregular (aperiodic) motion in which transported quantities (mass, momentum, scalar species) fluctuate in time and space z z Fluid properties and velocity exhibit random variations z z Identifiable swirling patterns characterize turbulent eddies. Enhanced mixing (matter, momentum, energy, etc.) results Statistical averaging results in accountable, turbulence related transport mechanisms. This characteristic allows for turbulence modeling. Contains a wide range of turbulent eddy sizes (scales spectrum). z The size/velocity of large eddies is on the order of mean flow. z Large eddies derive energy from the mean flow Energy is transferred from larger eddies to smaller eddies © 2006 ANSYS, Inc. All rights reserved. In the smallest eddies, turbulent energy is converted to internal energy by viscous dissipation. 6-2 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Is the Flow Turbulent? External Flows ρU L µ L = x, d , d h , etc. where Re L = Re x ≥ 500,000 along a surface Re d ≥ 20,000 around an obstacle Other factors such as free-stream turbulence, surface conditions, and disturbances may cause transition to turbulence at lower Reynolds numbers Internal Flows Re d h ≥ 2,300 Natural Convection 2 3 3 ρ C β g L ∆T β g L ∆ T Ra p 9 ≥ 10 = where Ra = is the Rayleigh number να µk Pr ν µCp Pr = = is the Prandtl number k α © 2006 ANSYS, Inc. All rights reserved. 6-3 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com Turbulent Flow Structures Small structures Large structures Energy Cascade Richardson (1922) © 2006 ANSYS, Inc. All rights reserved. 6-4 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Overview of Computational Approaches Reynolds-Averaged Navier-Stokes (RANS) models z z z Large Eddy Simulation (LES) z z Solves the spatially averaged N-S equations. Large eddies are directly resolved, but eddies smaller than the mesh are modeled. Less expensive than DNS, but the amount of computational resources and efforts are still too large for most practical applications. Direct Numerical Simulation (DNS) z z z Solve ensemble-averaged (or time-averaged) Navier-Stokes equations All turbulent length scales are modeled in RANS. The most widely used approach for calculating industrial flows. Theoretically, all turbulent flows can be simulated by numerically solving the full Navier-Stokes equations. Resolves the whole spectrum of scales. No modeling is required. But the cost is too prohibitive! Not practical for industrial flows - DNS is not available in Fluent. There is not yet a single, practical turbulence model that can reliably predict all turbulent flows with sufficient accuracy. © 2006 ANSYS, Inc. All rights reserved. 6-5 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Turbulence Models Available in FLUENT RANS based models © 2006 ANSYS, Inc. All rights reserved. One-Equation Models Spalart-Allmaras Two-Equation Models Standard k–ε RNG k–ε Realizable k–ε Standard k–ω SST k–ω Reynolds Stress Model Detached Eddy Simulation Large Eddy Simulation 6-6 Increase in Computational Cost Per Iteration ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com RANS Modeling – Time Averaging Ensemble (time) averaging may be used to extract the mean flow properties from the instantaneous ones: 1 ui (x, t ) = lim N →∞ N N ∑ ui (n ) (x, t ) ui′(x, t ) n =1 ui (x, t ) ui (x, t ) = ui (x, t ) + ui′(x, t ) Instantaneous component ui (x, t ) Example: Fully-Developed Turbulent Pipe Flow Velocity Profile Time-average Fluctuating component component The Reynolds-averaged momentum equations are as follows ∂u ∂u ∂ ∂p ρ i + uk i = − + ∂xk ∂xi ∂x j ∂t z ∂ui µ ∂x j ∂ Rij + ∂x j Rij = −ρui′u ′j (Reynolds stress tensor) The Reynolds stresses are additional unknowns introduced by the averaging procedure, hence they must be modeled (related to the averaged flow quantities) in order to close the system of governing equations. © 2006 ANSYS, Inc. All rights reserved. 6-7 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com The Closure Problem The RANS models can be closed in one of the following ways (1) Eddy Viscosity Models (via the Boussinesq hypothesis) ∂ui ∂u j 2 ∂uk 2 − µT + Rij = −ρ ui′u′j = µT δij − ρ k δij ∂x ∂x 3 ∂x 3 i k j z Boussinesq hypothesis – Reynolds stresses are modeled using an eddy (or turbulent) viscosity, µT. The hypothesis is reasonable for simple turbulent shear flows: boundary layers, round jets, mixing layers, channel flows, etc. (2) Reynolds-Stress Models (via transport equations for Reynolds stresses) z z Modeling is still required for many terms in the transport equations. RSM is more advantageous in complex 3D turbulent flows with large streamline curvature and swirl, but the model is more complex, computationally intensive, more difficult to converge than eddy viscosity models. © 2006 ANSYS, Inc. All rights reserved. 6-8 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Calculating Turbulent Viscosity Based on dimensional analysis, µT can be determined from a turbulence time scale (or velocity scale) and a length scale. z z z k = ui′ui′ 2 Turbulent kinetic energy [L2/T2] Turbulence dissipation rate [L2/T3] ε = ν ∂ui′ ∂x j (∂ui′ ∂x j + ∂u′j ∂xi ) Specific dissipation rate [1/T] ω=ε k Each turbulence model calculates µT differently. z Spalart-Allmaras: z Standard k–ε, RNG k–ε, Realizable k–ε z Solves a transport equation for a modified turbulent viscosity. Solves transport equations for k and ε. Standard k–ω, SST k–ω © 2006 ANSYS, Inc. All rights reserved. Solves transport equations for k and ω. 6-9 µT = f (~ ν) ρk2 µT = f ε ρk µT = f ω ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com The Spalart-Allmaras Model Spalart-Allmaras is a low-cost RANS model solving a transport equation for a modified eddy viscosity. z Mainly intended for aerodynamic/turbomachinery applications with mild separation, such as supersonic/transonic flows over airfoils, boundary-layer flows, etc. Embodies a relatively new class of one-equation models where it is not necessary to calculate a length scale related to the local shear layer thickness. Designed specifically for aerospace applications involving wall-bounded flows. z z When in modified form, the eddy viscosity is easy to resolve near the wall. Has been shown to give good results for boundary layers subjected to adverse pressure gradients. Gaining popularity for turbomachinery applications. This model is still relatively new. z z No claim is made regarding its applicability to all types of complex engineering flows. Cannot be relied upon to predict the decay of homogeneous, isotropic turbulence. © 2006 ANSYS, Inc. All rights reserved. 6-10 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com The k–ε Turbulence Models Standard k–ε (SKE) model z z z z The most widely-used engineering turbulence model for industrial applications Robust and reasonably accurate Contains submodels for compressibility, buoyancy, combustion, etc. Limitations The ε equation contains a term which cannot be calculated at the wall. Therefore, wall functions must be used. Generally performs poorly for flows with strong separation, large streamline curvature, and large pressure gradient. Renormalization group (RNG) k–ε model z z Constants in the k–ε equations are derived using renormalization group theory. Contains the following submodels z Differential viscosity model to account for low Re effects Analytically derived algebraic formula for turbulent Prandtl / Schmidt number Swirl modification Performs better than SKE for more complex shear flows, and flows with high strain rates, swirl, and separation. © 2006 ANSYS, Inc. All rights reserved. 6-11 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com The k–ε Turbulence Models Realizable k–ε (RKE) model z The term realizable means that the model satisfies certain mathematical constraints on the Reynolds stresses, consistent with the physics of turbulent flows. Positivity of normal stresses: ui′u ′j > 0 z z ( ) Schwarz’ inequality for Reynolds shear stresses: ui′u′j 2 ≤ ui2u 2j Neither the standard k–ε model nor the RNG k–ε model is realizable. Benefits: © 2006 ANSYS, Inc. All rights reserved. More accurately predicts the spreading rate of both planar and round jets. Also likely to provide superior performance for flows involving rotation, boundary layers under strong adverse pressure gradients, separation, and recirculation. 6-12 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com The k–ω Turbulence Models The k–ω family of turbulence models have gained popularity mainly because: z z The model equations do not contain terms which are undefined at the wall, i.e. they can be integrated to the wall without using wall functions. They are accurate and robust for a wide range of boundary layer flows with pressure gradient. FLUENT offers two varieties of k–ω models. z Standard k–ω (SKW) model z Most widely adopted in the aerospace and turbo-machinery communities. Several sub-models/options of k–ω: compressibility effects, transitional flows and shear-flow corrections. Shear Stress Transport k–ω (SSTKW) model (Menter, 1994) © 2006 ANSYS, Inc. All rights reserved. The SST k–ω model uses a blending function to gradually transition from the standard k–ω model near the wall to a high Reynolds number version of the k–ε model in the outer portion of the boundary layer. Contains a modified turbulent viscosity formulation to account for the transport effects of the principal turbulent shear stress. 6-13 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Large Eddy Simulationn Large Eddy Simulation (LES) z LES has been most successful for high-end applications where the RANS models fail to meet the needs. For example: Implementations in FLUENT: z Subgrid scale (SGS) turbulent models: z Combustion Mixing External Aerodynamics (flows around bluff bodies) Smagorinsky-Lilly model Wall-Adapting Local Eddy-Viscosity (WALE) Dynamic Smagorinsky-Lilly model Dynamic Kinetic Energy Transport Detached eddy simulation (DES) model LES is applicable to all combustion models in FLUENT Basic statistical tools are available: Time averaged and RMS values of solution variables, built-in fast Fourier transform (FFT). Before running LES, consult guidelines in the “Best Practices For LES” (containing advice for meshing, subgrid model, numerics, BCs, and more) © 2006 ANSYS, Inc. All rights reserved. 6-14 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Law of the Wall and Near-Wall Treatments Dimensionless velocity data from a wide variety of turbulent duct and boundary-layer flows are shown here: τw Wall shear Uτ = stress ρ u yUτ + + u = y = Uτ ν where y is the normal distance from the wall © 2006 ANSYS, Inc. All rights reserved. 6-15 For equilibrium turbulent boundary layers, walladjacent cells in the log-law region have known velocity and wall shear stress data ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com Wall Boundary Conditions The k–ε family and RSM models are not valid in the near-wall region, whereas Spalart-Allmaras and k–ω models are valid all the way to the wall (provided the mesh is sufficiently fine). To work around this, we can take one of two approaches. Wall Function Approach z z z z Standard wall function method is to take advantage of the fact that (for equilibrium turbulent boundary layers), a log-law correlation can supply the required wall boundary conditions (as illustrated in the previous slide). Non-equilibrium wall function method attempts to improve the results for flows with higher pressure gradients, separations, reattachment and stagnation. Similar laws are also constructed for the energy and species equations. Benefit: Wall functions allow the use of a relatively coarse mesh in the near-wall region. Enhanced Wall Treatment Option z z z z Combines a blended law-of-the wall and a two-layer zonal model. Suitable for low-Re flows or flows with complex near-wall phenomena Turbulence models are modified for the inner layer. Generally requires a fine near-wall mesh capable of resolving the viscous sublayer (at least 10 cells within the “inner layer”) © 2006 ANSYS, Inc. All rights reserved. 6-16 outer layer inner layer ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com Placement of The First Grid Point For standard or non-equilibrium wall functions, each wall-adjacent cell + centroid should be located within the log-law layer y p ≈ 30 − 300 For enhanced wall treatment (EWT), each wall-adjacent cell centroid should + be located within the viscous sublayer y p ≈ 1 z EWT can automatically accommodate cells placed in the log-law layer How to estimate the size of wall-adjacent cells before creating the grid: y +p = z y p uτ ν yp = Uτ = uτ τw Cf = Ue 2 ρ The skin friction coefficient can be estimated from empirical correlations: Flat plate: ⇒ y +p ν Cf 0.037 ≈ 2 Re1L 5 Duct: Cf 2 ≈ 0.039 Re1D4h Use postprocessing tools (XY plot or contour plot) to to double check the nearwall grid placement after the flow pattern has been established. © 2006 ANSYS, Inc. All rights reserved. 6-17 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Near-Wall Modeling: Recommended Strategy Use standard or non-equilibrium wall functions for most high Reynolds number applications (Re > 106) for which you cannot afford to resolve the viscous sublayer. z z You may consider using enhanced wall treatment if: z z There is little gain from resolving the viscous sublayer. The choice of core turbulence model is more important. Use non-equilibrium wall functions for mildly separating, reattaching, or impinging flows. The characteristic Reynolds number is low or if near wall characteristics need to be resolved. The physics and near-wall mesh of the case is such that y+ is likely to vary significantly over a wide portion of the wall region. Try to make the mesh either coarse or fine enough to avoid placing the wall-adjacent cells in the buffer layer (5 < y+ < 30). © 2006 ANSYS, Inc. All rights reserved. 6-18 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Inlet and Outlet Boundary Conditions When turbulent flow enters a domain at inlets or outlets (backflow), boundary conditions for k, ε, ω and/or ui′u ′j must be specified, depending on which turbulence model has been selected Four methods for directly or indirectly specifying turbulence parameters: z Explicitly input k, ε, ω, or z This is the only method that allows for profile definition. See user’s guide for the correct scaling relationships among them. Turbulence intensity and length scale Length scale is related to size of large eddies that contain most of energy. For boundary layer flows: l ≈ 0.4 δ99 For flows downstream of grid: l ≈ opening size z Turbulence intensity and hydraulic diameter z Turbulence intensity and turbulent viscosity ratio Ideally suited for internal (duct and pipe) flows For external flows: 1 < µt/µ < 10 Turbulence intensity depends on upstream conditions: I = u′ 1 ≈ U U 2k < 20 % 3 Stochastic inlet boundary conditions for LES and RANS can be generated by using spectral synthesizer or vortex method © 2006 ANSYS, Inc. All rights reserved. 6-19 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com GUI for Turbulence Models Define Models Viscous… Inviscid, Laminar, or Turbulent Define Boundary Conditions… Turbulence Model Options Near Wall Treatments Additional Options © 2006 ANSYS, Inc. All rights reserved. 6-20 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Example #1 – Turbulent Flow Past a Blunt Plate Reynolds-Stress model (“exact”) Standard k-ε model The Standard k–ε model is known to give spuriously large TKE on the font face of the plate Contour plots of turbulent kinetic energy (TKE) © 2006 ANSYS, Inc. All rights reserved. 6-21 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Example #1 – Turbulent Flow Past a Blunt Plate Skin Friction Coefficient Predicted separation bubble: Standard k-ε (SKE) Realizable k-ε (RKE) SKE severely underpredicts the size of the separation bubble, while RKE model predicts the size exactly. © 2006 ANSYS, Inc. All rights reserved. Experimentally observed reattachment point is at x/d = 4.7 6-22 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Example #2 – Turbulent Flow in a Cyclone 0.1 m 40,000-cell hexahedral mesh High-order upwind scheme was used. Computed using SKE, RNG, RKE and RSM (second moment closure) models with the standard wall functions Represents highly swirling flows (Wmax = 1.8 Uin) 0.12 m Uin = 20 m/s 0.97 m 0.2 m © 2006 ANSYS, Inc. All rights reserved. 6-23 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Example #2 – Turbulent Flow in a Cyclone Tangential velocity profile predictions at 0.41 m below the vortex finder © 2006 ANSYS, Inc. All rights reserved. 6-24 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com Example #3 – Flow Past a Square Cylinder (LES) Drag Coefficient Strouhal Number Dynamic Smagorinsky 2.28 0.130 Dynamic TKE 2.22 0.134 2.1 – 2.2 0.130 Exp.(Lyn et al., 1992) Time-averaged streamwise velocity along the wake centerline Iso-Contours of Instantaneous Vorticity Magnitude © 2006 ANSYS, Inc. All rights reserved. (ReH = 22,000) CL spectrum 6-25 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Example #3 – Flow Past a Square Cylinder (LES) Streamwise mean velocity along the wake centerline © 2006 ANSYS, Inc. All rights reserved. Streamwise normal stress along the wake centerline 6-26 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Summary: Turbulence Modeling Guidelines Successful turbulence modeling requires engineering judgment of: z z z Flow physics Computer resources available Project requirements z Accuracy Turnaround time Near-wall treatments Modeling procedure z z z z z Calculate characteristic Re and determine whether the flow is turbulent. Estimate wall-adjacent cell centroid y+ before generating the mesh. Begin with SKE (standard k-ε) and change to RKE, RNG, SKW, SST or V2F if needed. Check the tables in the appendix as a starting guide. Use RSM for highly swirling, 3-D, rotating flows. Use wall functions for wall boundary conditions except for the low-Re flows and/or flows with complex near-wall physics. © 2006 ANSYS, Inc. All rights reserved. 6-27 ANSYS, Inc. Proprietary Appendix © 2006 ANSYS, Inc. All rights reserved. ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com RANS Turbulence Model Descriptions Model Description Spalart – Allmaras A single transport equation model solving directly for a modified turbulent viscosity. Designed specifically for aerospace applications involving wall-bounded flows on a fine near-wall mesh. FLUENT’s implementation allows the use of coarser meshes. Option to include strain rate in k production term improves predictions of vortical flows. Standard k–ε The baseline two-transport-equation model solving for k and ε. This is the default k–ε model. Coefficients are empirically derived; valid for fully turbulent flows only. Options to account for viscous heating, buoyancy, and compressibility are shared with other k–ε models. RNG k–ε A variant of the standard k–ε model. Equations and coefficients are analytically derived. Significant changes in the ε equation improves the ability to model highly strained flows. Additional options aid in predicting swirling and low Reynolds number flows. Realizable k–ε A variant of the standard k–ε model. Its “realizability” stems from changes that allow certain mathematical constraints to be obeyed which ultimately improves the performance of this model. Standard k–ω A two-transport-equation model solving for k and ω, the specific dissipation rate (ε / k) based on Wilcox (1998). This is the default k–ω model. Demonstrates superior performance for wall-bounded and low Reynolds number flows. Shows potential for predicting transition. Options account for transitional, free shear, and compressible flows. SST k–ω A variant of the standard k–ω model. Combines the original Wilcox model for use near walls and the standard k–ε model away from walls using a blending function. Also limits turbulent viscosity to guarantee that τT ~ k. The transition and shearing options are borrowed from standard k–ω. No option to include compressibility. Reynolds Stress Reynolds stresses are solved directly using transport equations, avoiding isotropic viscosity assumption of other models. Use for highly swirling flows. Quadratic pressure-strain option improves performance for many basic shear flows. © 2006 ANSYS, Inc. All rights reserved. 6-29 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com RANS Turbulence Model Behavior and Usage Model Behavior and Usage Spalart-Allmaras Economical for large meshes. Performs poorly for 3D flows, free shear flows, flows with strong separation. Suitable for mildly complex (quasi-2D) external/internal flows and boundary layer flows under pressure gradient (e.g. airfoils, wings, airplane fuselages, missiles, ship hulls). Standard k–ε Robust. Widely used despite the known limitations of the model. Performs poorly for complex flows involving severe pressure gradient, separation, strong streamline curvature. Suitable for initial iterations, initial screening of alternative designs, and parametric studies. RNG k–ε Suitable for complex shear flows involving rapid strain, moderate swirl, vortices, and locally transitional flows (e.g. boundary layer separation, massive separation, and vortex shedding behind bluff bodies, stall in wide-angle diffusers, room ventilation). Realizable k–ε Offers largely the same benefits and has similar applications as RNG. Possibly more accurate and easier to converge than RNG. Standard k–ω Superior performance for wall-bounded boundary layer, free shear, and low Reynolds number flows. Suitable for complex boundary layer flows under adverse pressure gradient and separation (external aerodynamics and turbomachinery). Can be used for transitional flows (though tends to predict early transition). Separation is typically predicted to be excessive and early. SST k–ω Offers similar benefits as standard k–ω. Dependency on wall distance makes this less suitable for free shear flows. Reynolds Stress Physically the most sound RANS model. Avoids isotropic eddy viscosity assumption. More CPU time and memory required. Tougher to converge due to close coupling of equations. Suitable for complex 3D flows with strong streamline curvature, strong swirl/rotation (e.g. curved duct, rotating flow passages, swirl combustors with very large inlet swirl, cyclones). © 2006 ANSYS, Inc. All rights reserved. 6-30 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com The Spalart-Allmaras Turbulence Model A low-cost RANS model solving an equation for the modified eddy viscosity, ~ν 2 ~ ~ ~ Dν ∂ ν ∂ν ∂ ~ ( ) C = Gν µ + ρ ν + ρ − Yν + S ~ν b2 Dt ∂x j ∂x j ∂x j Eddy viscosity is obtained from µ = ρ~ νf t 3 ~ ( ν / ν) f v1 = ~ 3 (ν / ν ) + Cv31 v1 The variation of ~ ν very near the wall is easier to resolve than k and ε. Mainly intended for aerodynamic/turbomachinery applications with mild separation, such as supersonic/transonic flows over airfoils, boundary-layer flows, etc. © 2006 ANSYS, Inc. All rights reserved. 6-31 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com RANS Models - Standard k–ε (SKE) Model Transport equations for k and ε: D ∂ (ρ k ) = µ + µ t Dt ∂x j σk ∂k + Gk − ρ ε ∂x j µ t ∂ε D ∂ ε ε2 (ρ ε ) = µ + + Ce1 Gk − ρ Cε 2 Dt k k ∂x j σ ε ∂x j where Cµ = 0.09, Cε1 = 1.44, Cε 2 = 1.92, σ k = 1.0, σ ε = 1.3 The most widely-used engineering turbulence model for industrial applications Robust and reasonably accurate; it has many sub-models for compressibility, buoyancy, and combustion, etc. Performs poorly for flows with strong separation, large streamline curvature, and high pressure gradient. © 2006 ANSYS, Inc. All rights reserved. 6-32 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com RANS Models – k–ω Models µt = α∗ ρ ρ ρ k ω ∂u Dk ∂ = τij i − ρ β∗ f β∗ k ω + Dt ∂x j ∂x j specific dissipation rate, ω µt µ + σk ω ∂u ∂ Dω = α τij i − ρ β f β ω2 + ∂x j Dt k ∂x j ∂k x ∂ j µ t ∂ω µ + σ ω ∂x j ε 1 ω≈ ∝ k τ Belongs to the general 2-equation EVM family. Fluent 6 supports the standard k–ω model by Wilcox (1998), and Menter’s SST k–ω model (1994). k–ω models have gained popularity mainly because: z Can be integrated to the wall without using any damping functions z Accurate and robust for a wide range of boundary layer flows with pressure gradient Most widely adopted in the aerospace and turbo-machinery communities. Several sub-models/options of k–ω: compressibility effects, transitional flows and shear-flow corrections. © 2006 ANSYS, Inc. All rights reserved. 6-33 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com RANS Models – Reynolds Stress Model (RSM) ( ) ( ) ∂ ∂ ρui′u ′j + ρ u k ui′u ′j = Pij + Fij + DijT + Φ ij − ε ij ∂t ∂xk Stress production Rotation production Dissipation Turbulent diffusion Pressure Strain Modeling required for these terms Attempts to address the deficiencies of the EVM. RSM is the most ‘physically sound’ model: anisotropy, history effects and transport of Reynolds stresses are directly accounted for. RSM requires substantially more modeling for the governing equations (the pressure-strain is most critical and difficult one among them). But RSM is more costly and difficult to converge than the 2-equation models. Most suitable for complex 3-D flows with strong streamline curvature, swirl and rotation. © 2006 ANSYS, Inc. All rights reserved. 6-34 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com Standard Wall Functions Standard Wall Functions z Momentum boundary condition based on Launder-Spaulding law-of-thewall: y∗ U∗ = 1 ∗ ln E y κ ( z z z ) for y∗ < yν∗ ∗ ∗ ν for y > y ∗ where U = U P Cµ1/ 4 kP1/ 2 U 2 τ y∗ = ρ Cµ1/ 4 kP1/ 2 y p µ Similar wall functions apply for energy and species. Additional formulas account for k, ε, and ρ ui′u′j . Less reliable when flow departs from conditions assumed in their derivation. © 2006 ANSYS, Inc. All rights reserved. Severe ∇p or highly non-equilibrium near-wall flows, high transpiration or body forces, low Re or highly 3D flows 6-35 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com Standard Wall Functions Energy ρ Cµ1 4 k1 2 Prt U P2 + (Pr− Prt )Uc2 2q T∗ = 14 12 2 ρ Pr C k U 1 1 µ P ∗ ∗ ∗ Pr ln Ey + P + Pr y + ln E y t κ κ 2q [ ] ( ) ( for y∗ < yt∗ ) for y∗ > yt∗ Pr 3 4 Pr P = 9.24 −1 1 + 0.28exp − 0.007 Prt Prt Species Sc y∗ for y∗ < yc∗ Y∗ = 1 ∗ ∗ ∗ + > E y P y y Sc ln for t c c κ ( © 2006 ANSYS, Inc. All rights reserved. ) 6-36 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Non-Equilibrium Wall Functions Non-equilibrium wall functions z Standard wall functions are modified to account for stronger pressure gradients and non-equilibrium flows. Useful for mildly separating, reattaching, or impinging flows. Less reliable for high transpiration or body forces, low Re or highly 3D flows. The standard and non-equilibrium wall functions are options for all of the k–ε models as well as the Reynolds stress model. © 2006 ANSYS, Inc. All rights reserved. 6-37 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com Enhanced Wall Treatment Enhanced wall functions z z z Momentum boundary condition based on u + = e Γ u + + u + e1 Γ lam turb a blended law-of-the-wall (Kader). Similar blended wall functions apply for energy, species, and ω. Kader’s form for blending allows for incorporation of additional physics. Pressure gradient effects Thermal (including compressibility) effects Two-layer zonal model z A blended two-layer model is used to determine near-wall ε field. z Domain is divided into viscosity-affected (near-wall) region and turbulent core region. ρy k Based on the wall-distance-based turbulent Reynolds number: Re y ≡ µ Zoning is dynamic and solution adaptive. High Re turbulence model used in outer layer. Simple turbulence model used in inner layer. Solutions for ε and µT in each region are blended: λ ε (µ t )outer + (1 − λ ε )(µ t )inner The Enhanced Wall Treatment option is available for the k–ε and RSM models (EWT is the sole treatment for Spalart Allmaras and k–ω models). © 2006 ANSYS, Inc. All rights reserved. 6-38 ANSYS, Inc. Proprietary Fluent User Services Center Introductory FLUENT Notes FLUENT v6.3 December 2006 www.fluentusers.com Two-Layer Zonal Model The two regions are demarcated on a cell-by-cell basis: z z Turbulent core region Re y > 200 Viscosity affected region Re y < 200 z z y is the distance to the nearest wall Zoning is dynamic and solution adaptive Re y ≡ © 2006 ANSYS, Inc. All rights reserved. 6-39 ρy k µ ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Turbulent Heat Transfer The Reynolds averaging of the energy equation produces an additional term ui′t ′ z Analogous to the Reynolds stresses, this is the turbulent heat flux term. An isotropic turbulent diffusivity is assumed: ui′t ′ = − νT z Turbulent diffusivity is usually related to eddy viscosity via a turbulent Prandtl number (modifiable by the users): Prt = ∂T ∂xi νt ≈ 0.85 − 0.9 νT Similar treatment is applicable to other turbulent scalar transport equations. © 2006 ANSYS, Inc. All rights reserved. 6-40 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Menter’s SST k–ω Model Background Many people, including Menter (1994), have noted that: z z z The k–ω model has many good attributes and performs much better than k–ε models for boundary layer flows. Wilcox’ original k–ω model is overly sensitive to the free stream value of ω, while the k–ε model is not prone to such problem. Most two-equation models, including k–ε models, over predict turbulent stresses in wake (velocity-defect) regions, which leads to poor performance of the models for boundary layers under adverse pressure gradient and separated flows. © 2006 ANSYS, Inc. All rights reserved. 6-41 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Menter’s SST k–ω Model Main Components The SST k–ω model consists of z z Zonal (blended) k–ω / k–ε equations (to address item 1 and 2 in the previous slide) Clipping of turbulent viscosity so that turbulent stress stay within what is dictated by the structural similarity constant. (Bradshaw, 1967) - addresses item 3 in the previous slide Outer layer (wake and outward) Inner layer (sub-layer, log-layer) k–ω model transformed from standard k–ε model 3 k 2 ε= l Modified Wilcox’ε k–ω model Wilcox’ original k-ω model Wall © 2006 ANSYS, Inc. All rights reserved. 6-42 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Menter’s SST k–ω Model Blended equations The resulting blended equations are: µT ∂k µ + σ k ∂x j ∂ui ∂ µT ∂ω Dω γ 1 ∂k ∂ω 2 ( ) ρ = τij − βρω + µ + + ρ − F σ 2 1 1 ω2 ∂x j σ ω ∂x j Dt ν t ∂x j ω ∂x j ∂x j ∂u ∂ Dk ρ = τij i − β* k ρ ω + ∂x j ∂x j Dt φ = F1 φ1 + (1 − F1 ) φ1 ; φ = β, σ k , σ ω , γ Wall © 2006 ANSYS, Inc. All rights reserved. 6-43 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Large Eddy Simulation (LES) N-S equation ∂ui ∂ui u j 1 ∂p ∂ =− + + ∂x j ρ ∂xi ∂x j ∂t ∂ui ν ∂x j ui (x, t ) = ui (x, t ) + ui′(x, t ) Instantaneous component Resolved Scale Subgrid Scale Filtered N-S equation ∂ui ∂ui u j 1 ∂p ∂ + =− + ∂t ∂x j ρ ∂xi ∂x j ∂ui ν ∂x j ∂ τij − ∂x j Filter, ∆ ( τij = ρ ui u j − ui u j ) (Subgrid scale Turbulent stress) Spectrum of turbulent eddies in the Navier-Stokes equations is filtered: z z z The filter is a function of grid size Eddies smaller than the grid size are removed and modeled by a subgrid scale (SGS) model. Larger eddies are directly solved numerically by the filtered transient N-S equation © 2006 ANSYS, Inc. All rights reserved. 6-44 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com LES in FLUENT LES has been most successful for high-end applications where the RANS models fail to meet the needs. For example: z z z Combustion Mixing External Aerodynamics (flows around bluff bodies) Implementations in FLUENT: z Sub-grid scale (SGS) turbulent models: z Smagorinsky-Lilly model WALE model Dynamic Smagorinsky-Lilly model Dynamic kinetic energy transport model Detached eddy simulation (DES) model LES is applicable to all combustion models in FLUENT Basic statistical tools are available: Time averaged and RMS values of solution variables, built-in fast Fourier transform (FFT). Before running LES, consult guidelines in the “Best Practices For LES” (containing advice for meshing, subgrid model, numerics, BCs, and more) © 2006 ANSYS, Inc. All rights reserved. 6-45 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Detached Eddy Simulation (DES) Motivation z z For high-Re wall bounded flows, LES becomes prohibitively expensive to resolve the near-wall region Using RANS in near-wall regions would significantly mitigate the mesh resolution requirement RANS/LES hybrid model based on the Spalart-Allmaras turbulence model: 2 D~ 1 ∂ ν ν ~~ ~ = Cb1 S ν − Cw1 f w + Dt d σ ~ν ∂x j ∂~ ν ~ (µ + ρ ν ) + ... ∂x j d = min (d w , CDES∆ ) z z One-equation SGS turbulence model In equilibrium, it reduces to an algebraic model. DES is a practical alternative to LES for high-Reynolds number flows in external aerodynamic applications © 2006 ANSYS, Inc. All rights reserved. 6-46 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com V2F Turbulence Model A model developed by Paul Durbin’s group at Stanford University. z z z Durbin suggests that the wall-normal fluctuations v′2 are responsible for the near-wall damping of the eddy viscosity Requires two additional transport equations for v′2 and a relaxation function f to be solved together with k and ε. Eddy viscosity model is ν T ~ v′2 T instead of νT ~ k T V2F shows promising results for many 3D, low Re, boundary layer flows. For example, improved predictions for heat transfer in jet impingement and separated flows, where k–ε models fail. But V2F is still an eddy viscosity model and thus the same limitations still apply. V2F is an embedded add-on functionality in FLUENT which requires a separate license from Cascade Technologies (www.turbulentflow.com) © 2006 ANSYS, Inc. All rights reserved. 6-47 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Stochastic Inlet Velocity Boundary Condition It is often important to specify realistic turbulent inflow velocity BC for accurate prediction of the downstream flow: ui (x, t ) = ui (x ) + ui′(x, t ) Instantaneous Coherent + Random Different types of inlet boundary conditions for LES z z z No perturbations – Turbulent fluctuations are not present at the inlet. Vortex method – Turbulence is mimicked by using the velocity field induced by many quasi-random point-vortices on the inlet surface. The vortex method uses turbulence quantities as input values (similar to those used for RANS-based models). Spectral synthesizer Time Averaged Able to synthesize anisotropic, inhomogeneous turbulence from RANS results (k–ε, k– ω, and RSM fields). Can be used for RANS/LES zonal hybrid approach © 2006 ANSYS, Inc. All rights reserved. 6-48 ANSYS, Inc. Proprietary Introductory FLUENT Notes FLUENT v6.3 December 2006 Fluent User Services Center www.fluentusers.com Initial Velocity Field for LES/DES Initial condition for velocity field does not affect statistically stationary solutions However, starting LES with a realistic turbulent velocity field can substantially shorten the simulation time to get to statistically stationary state The spectral synthesizer can be used to superimpose turbulent velocity on top of the mean velocity field z z Uses steady-state RANS (k–ε, k–ω, RSM, etc.) solutions as inputs to the spectral synthesizer Accessible via a TUI command: /solve/initialize/init-instantaneous-vel © 2006 ANSYS, Inc. All rights reserved. 6-49 ANSYS, Inc. Proprietary